The pryout capacity is based on the tension breakout capacity, so if you design the supplemental reinforcement as tension breakout reinforcement, then that should cover you. Or if the pryout capacity is greater than the loads, you can leave as-is. But a few pieces of rebar are probably cheap insurance in any case.

I don't think the code explicitly covers this, but pryout is inherently a tension cone failure, so it makes sense to design the supplemental reinforcement on that basis. But I think horizontal hairpins at the anchor head could also prevent pryout, since the anchor head becomes restrained from rotating and popping the concrete up behind it. It's kind of a judgment call.

Good points bones.

1) A scan of ACI 318-19 doesn't yield any silver bullet that you might use.

a) For supplementary reinforcement, it's set up in such a way that it's pretty clear that they don't intend for that to be used for pry out.

b) For true anchorage reinforcement, the only make specific reference to shear breakout and tension breakout.

2) A PCI paper on pryout indicates that it's not an issue for hef/d < 4.5 which is consistent with what you mentioned. That said, they don't really say anything about the case of rear anchors near a free edge.

3) The rear anchor cracks shown below are suggestive of your rear edge distance concern.

4) I really do think that the character of this changes substantially with long-ish anchors. I'll noodle some on a) finding a way to justify no prying action and b) the best way to reinforce for prying action given expected behavior.

Thanks Kootk. Those studs look very short.

The thing I have to wonder is: if I have good bearing on the side of the bolt (with a pretty conservative model).....and 8d embedment......can it really pry out (even with that close edge on the back end)? Will read on this some more myself.

The calculation of group tension breakout capacity should account for the influence of the near edge. So it’s not like the existing payout equation does not account for it at all.

Quote (bones206)

The calculation of group tension breakout capacity should account for the influence of the near edge.

It's tempting to think that but, then, look at the pictures above. Clearly, at some degree of separation, you produce separate, and apparently different, pryout phenomena.

Quote (OP)

Those studs look very short.

No doubt. Fundamentally, the source of pryout is the eccentricity between where the shear is delivered (plate) and where it is resisted (a little below the plate). So the general tendency is there with any proportion of anchor. With the short and stocky, we seem to understand the mechanism. The anchor does kind of a rigid body rotation along with the up and out diagonal kickback thing. With a long anchor, I'm not sure that I know the mechanism. I sketched up a few things but didn't have any real confidence in any of them. Regardless, with a longer, skinnier anchor, I'm confident that your prying action exhibits the following features:

1) A greater lever arm for the concrete compression stresses that ultimately resist the applied eccentricity. Ergo much smaller resisting couple loads.

2) Pry out demand that tends more to wards kicking out than kicking up and out.

Both of the those things help but, of course, the questions is how much. #2 leads me to believe that hairpins parallel to load are the way to go for this. But then I'm unsure where to place them best perpendicular to the load. maybe just add a pair at 3" oc near the bottom of the anchor and call it a day.

Further to bones' point, 318 indicates not pryout check for individual anchors within a group of anchors. The intent there seems pretty clear.

I guess what I'm saying is that pryout capacity near an edge will calc out to be less than pryout capacity in the middle of a slab. Just because the tension breakout area is reduced. I don't know how closely that reduction estimates the true pryout-near-an-edge effect, but at least there is some reduction in capacity. If the "breakout area method" is an acceptable estimate for tension breakout near an edge, I don't see why it wouldn't also be a good estimate of pryout near an edge.